from transformers import AutoModelForTokenClassification, AutoTokenizer
from fastapi import FastAPI
from fastapi.staticfiles import StaticFiles
from fastapi.middleware.cors import CORSMiddleware

app = FastAPI()

# Setup middlewares
app.add_middleware(
    CORSMiddleware,
    allow_origins=["*"],
    allow_credentials=True,
    allow_methods=["*"],
    allow_headers=["*"],
)

# Setup static files
app.mount("/static", StaticFiles(directory="static"), name="static")

@app.command("help")


@app.get("/ping")
async def ping():
    return {"status": "pong"}

@app.post("/predict")
async def predict(input_text: str):
    tokenizer = AutoTokenizer.from_pretrained("your_model_name")
    model = AutoModelForTokenClassification.from_pretrained("your_model_name")
    
    inputs = tokenizer([input_text], return_tensors="pt", padding=True, truncation=True)
    outputs = model(**inputs)
    prediction = outputs.logits.argmax().item()
    return {"prediction": prediction}